253 features
Music licensing is one of those problems everyone agrees is broken and almost no one has actually fixed, so I was curious whether ElevenLabs had something real here or just a rebrand of ambition.
The dictation app space is crowded with tools that almost work, and Voicr is making a case that almost isn't good enough.
ProdShort captures your real meeting moments and transforms them into short-form social content—no scripting, no AI voice, just your actual words.
Every DAW promises freedom and delivers a spreadsheet. Riffle is betting that the problem isn't your skills, it's the cage.
The fitness tracking app comparison is so worn out it's almost a red flag now, but Morsel might actually have a case.
Most identity verification checks if a document looks legit. PassportReader checks if it actually is, and that distinction is more important than the industry currently admits.
A tool that fills giant letters with tiny letters sounds like a novelty, but I've been thinking about it more than I expected to.
Every performance marketer I know is drowning in dashboards, and Predflow thinks the answer is one AI agent that just tells you what to do.
Most break reminder apps treat RSI like a footnote. Zzzappy built the whole product around it.
Everyone is trying to file taxes with a chatbot. Jupid is the first one that admits why that keeps failing.
Every Claude Code power user has hit the wall. Edgee thinks a token compressor running at the edge is the fix, and their benchmark numbers are harder to dismiss than I expected.
Bujo AI builds voice and chat agents for sales and customer support at large B2C enterprises, founded by a two-time YC founder who previously built self-driving car technology that was acquired.
Korso automates the full quote-to-order workflow for manufacturers, processing incoming RFQs from email and WhatsApp, extracting line items, generating professional quotes, and tracking purchase orders through production.
Replicas lets engineering teams delegate tasks to AI coding agents that run in sandboxed VMs with full development environments, accessible from GitHub, Slack, Linear, or a web dashboard, and reportedly shipping over 30% of pull requests for some teams.
Arga Labs provides sandbox infrastructure with fully functional mocks of external services so you can validate code changes and AI agent behavior at scale without risking production data or hitting real APIs.
Lamina Labs builds Pictor, an AI video generation tool that creates mathematically precise educational animations from a text prompt in under eight seconds, with zero hallucinations and zero approximations.
Zolvo automates reconciliation, verification, collections, and reporting for commercial lenders, replacing the spreadsheets and legacy software that back-office teams use to service billions in loan portfolios.
Arcline provides AI-native legal services for startups with same-day turnaround, using AI to handle 80% of the work so founders pay elite lawyers only for the final 20% that requires human judgment.
Cumulus Labs is a serverless GPU inference provider claiming 12.5-second cold starts and scale-to-zero pricing, positioning itself as faster than Modal and cheaper than RunPod for teams running AI models in production.
Hex Security builds AI agents that run continuous penetration tests against your apps and infrastructure, replacing the once-a-year pentest with an always-on attack simulation that finds vulnerabilities before real attackers do.
o11 integrates directly inside Microsoft PowerPoint, Excel, Word, and their Google equivalents, letting users automate workflows from slide decks to financial models without ever leaving the application they are already working in.
Space research has always been limited by the number of humans who can physically be there. General Astronautics builds autonomous robots for microgravity research and manufacturing, backed by YC and NVIDIA Inception.
MochaCare combines AI tools with human virtual assistants to run hiring, scheduling, and client intake around the clock for home care agencies that are too busy keeping patients alive to keep up with their own back office.
Recruiters spend 43.5 hours per month coordinating interviews across candidates, hiring managers, and panelists. Vela handles the scheduling, follow-ups, and timezone juggling across email, SMS, WhatsApp, and Slack.
Mortgage origination is drowning in paperwork. Copperlane's AI agent Penny handles document collection, verification, and borrower communication 24/7, so loan officers can focus on closing deals.
Fish hatcheries inspect every fish by hand, spending 5 minutes per specimen. OctaPulse does it in 30 seconds with 95%+ accuracy and has already signed a six-figure pilot with the largest trout producer in the US.
RunAnywhere provides an SDK for deploying AI models on iOS, Android, and edge devices. Their MetalRT engine runs 1.67x faster than llama.cpp on Apple Silicon. 10K+ GitHub stars and growing.
B2B SaaS companies know partnerships matter but cannot prove it with numbers. Arzule uses probabilistic models to quantify how much revenue flows through partner relationships, shared customers, and co-selling.
Ashr generates synthetic user journeys across voice, text, images, and tool calls to test AI agents before they hit production. Prompt versioning with inline diffs shows you exactly what changed and why it broke.
Commercial insurance still runs on a human API of brokers chasing email threads and re-keying data. Panta uses AI agents to automate the entire back-office, delivering 10x faster results with a 99% placement rate across 100+ carriers.
Most AI training data has clear right answers. Traverse focuses on the hard part: capturing how humans make subjective decisions in law, healthcare, sales, and writing to train models that need judgment.
Electrical contractors spend hours counting outlets, switches, and fixtures on floor plans by hand. Bidflow does it with computer vision at 95-99% accuracy, 20x faster, and 70% cheaper.
Grade lets companies pay any creator worldwide with just their email address. No onboarding required. 790+ creators paid, $1M+ in total payouts, and 120% month-over-month growth.
Martini replaces prompt-only workflows with real camera controls for AI video generation. Position your camera, pick your lens, compose your shot. Over 200 films created, with content airing as TV commercials.
Lexius adds an AI layer to existing security cameras that detects shoplifting, slip-and-fall incidents, and known offenders in real time. No new hardware. Clients include 7-Eleven, Erewhon, and Prada.
Librar Labs is building self-healing data infrastructure that turns fragmented library records into agent-ready systems. Their school library product reportedly doubles student reading rates.
Oxus automates the tedious parts of internal auditing for pre-IPO and public companies. Scoping, flowcharts, control testing, and workpaper generation that used to take days now takes minutes.
IncidentFox autonomously investigates production incidents by querying logs, metrics, and deployment history, delivering root cause analysis and one-click remediation scripts that wait for human approval.
ClaimGlide automates the entire prior authorization process for medical practices, from initial submission to payer follow-up to denial appeals, including AI-generated appeal letters and automated phone calls to payers.
Polymath builds complex multi-tool environments for training AI agents through reinforcement learning. Their Horizon-SWE benchmark tests what happens when you ask AI to do real work across real tools over long time horizons.
Wayco automates medical case management for personal injury law firms, using AI to match plaintiffs with the right medical providers, analyze records, and accelerate settlement timelines.
Fed10 monitors legislation across all 50 states and Congress, analyzes bills for business impact, and delivers alerts and policy intelligence at a fraction of the cost of traditional policy consultants.
Glue is an open source design canvas that integrates with coding agents like Claude Code and OpenAI Codex, letting developers leave visual feedback on their app and watch agents convert it to code in real time.
Talking Computers provides Facility, a workplace platform where enterprises deploy fully autonomous AI organizations with workers that collaborate with each other and human managers to get work done around the clock.
Ditto Bio uses AI to identify proteins from viruses, ticks, and worms that naturally suppress immune responses, then engineers them into therapeutics for rheumatoid arthritis, IBD, MS, and type 1 diabetes.
Pollen monitors communications, support tickets, product usage, and CRM data to detect back-tested churn and upsell signals, then tells your team which accounts need attention today.
Shofo aggregates public and private video sources into a continuously updating index of billions of videos, then cleans, searches, and labels them to build custom datasets for AI research labs.
Revion builds voice-first AI tools that automate repair order documentation, warranty compliance, and inspection workflows so auto technicians can spend more time turning wrenches and less time typing.
Skillsync creates structured skill profiles from developers' open source contributions on GitHub, letting companies search for specific capabilities like 'deployed custom neural nets on edge' and find engineers already building in their domain.
Terranox is the first vertically integrated AI-powered uranium exploration company, using machine learning trained on seven decades of exploration outcomes to find high-grade deposits in North America.
Velum Labs monitors production query traffic to detect when teams calculate the same metric differently, traces root causes through query lineage, and generates enforceable data contracts to prevent future breaks.
Kita automates credit review for lenders in emerging markets like the Philippines, where borrowers have no credit scores and financial history lives in unstructured documents.
Unifold provides a developer-first API and SDK that lets any app accept on-chain deposits across Ethereum, Solana, Bitcoin, and every EVM chain with less than 10 lines of code.
Valgo builds probabilistic risk models for autonomous systems so insurers can price coverage for robots, autonomous trucks, and drones that have nearly zero historical claims data.
Beacon Health builds AI workers that plug directly into EHR systems to handle risk adjustment, preventive screenings, and prior authorizations so primary care physicians can focus on patients.
Compresr provides an API that compresses LLM context up to 200x without losing what matters, cutting token costs and improving accuracy for agents and RAG systems.
Piris Labs offers a full-stack inference service using proprietary photonic hardware to eliminate the AI data movement bottleneck, delivering high-performance inference at a fraction of GPU cluster costs.
AxionOrbital builds foundation models that translate raw radar satellite data into optical-quality imagery, enabling continuous visibility through clouds and darkness.
Origin Bio designs AI-generated regulatory DNA sequences that program precise therapeutic gene expression patterns, targeting diseases like cancer and CNS disorders.
Sentrial is the monitoring platform for AI products in production, detecting real-time when agents hallucinate, loop, or frustrate users, then diagnosing root causes and recommending fixes.
Cardinal is the AI platform for precision outbound, running outbound sales operations for 40+ YC companies by finding genuine commonalities between sellers and prospects.
Bubble Lab deploys Pearl, an AI agent that lives in Slack, connects to your SaaS tools, and handles the ops work your team keeps putting off.
Cajal builds superhuman AI mathematicians that discover and verify proofs using Lean, targeting quantum computing and finance where mathematical certainty is not optional.
Balance is building an AI accounting firm for small businesses that reconciles books in real time and flags issues before they snowball into expensive problems.
Two UPenn friends who spent over a decade at NASA and SpaceX are building deployable solar arrays that unfold to massive scale once they reach orbit.
Healthcare data is fragmented across dozens of systems that do not talk to each other. Eos AI harmonizes it all into something clinics can actually use.
Building simulations used to take weeks. Inviscid AI does them in seconds with physics-informed neural networks. The energy savings alone make this a no-brainer for large facilities.
Veriad wants to replace the humans who check whether your ads are legal and on-brand. Their AI agents do it in under 20 seconds.
Video editing is the bottleneck nobody wants to talk about. Cardboard just threw an AI agent at it, and the results are surprisingly competent.
Personalized messaging is supposed to be easy by now. It is not. Polymorph builds living user profiles that actually drive 3.6x conversion increases, and the numbers are hard to argue with.
Most customer support waits for tickets. Return Signals texts customers after delivery, and 61% of them text back. Those are not normal response rates.
I have been waiting for someone to build the air traffic control tower for coding agents. Emdash is that tower, and it is open source.
While Ethereum DeFi keeps recycling the same ideas with different logos, Navi Protocol quietly became the dominant lending platform on one of crypto's fastest-growing chains.
Four Swedish founders looked at the patent industry and saw attorneys manually reading thousands of documents. They built an AI system that does it faster and shows its work.
QR codes are everywhere and they are all ugly. Glyph uses AI to generate QR codes that double as art, and the 5,500 images already created suggest the market agrees that this is a problem worth solving.
Creative automation for marketing teams sounds like the future until you realize how many startups have said that exact sentence. Pulp is early, the landing page is thin, and the question is whether the product behind it justifies the ambition.
The AI customer support space is brutally crowded. Sona is betting that e-commerce specificity and high resolution rates will carve out enough room to matter. The numbers are bold. The question is whether they hold up under real volume.
AI agents are everywhere and nobody can tell you why they broke. Fernstone gives developers tracing, evals, and cost tracking for autonomous AI systems so you can actually debug the thing before your customer does.
Chat interfaces had a good run. Cove is betting that AI collaboration works better on a canvas where you can see your thinking laid out spatially, not buried in a scrolling thread.
Redapto monitors customer support interactions across chat, voice, and email, detects quality issues, and automatically updates training materials and knowledge bases. Built by ex-Coinbase and ex-Amazon AI engineers.
Your CI pipeline is rebuilding code that has not changed, and you are paying for every wasted minute. Trace Machina's NativeLink is an open-source remote execution and caching system that handles over a billion requests a month for companies like Samsung.
Insurance claims adjusting is slow, manual, and expensive. Willow Labs is building AI agents that investigate claims, analyze evidence, and recommend payouts so adjusters can handle ten times the caseload.
Patent Watch automates the miserable work of tracking competitor patent filings. Upload your portfolio, and their AI scans 160 million patents to find who might be stepping on your claims.
Finance teams spend half their time logging into clunky ERP systems and doing the same reconciliation tasks over and over. Zalos deploys AI agents that do the clicking, the matching, and the uploading so humans can stop being data-entry robots.
If you have ever tried to run three coding agents in parallel on the same machine, you have hit the port clash wall. Coasts fixes that with localhost isolation per git worktree, and it is free.
Duolingo turned language learning into a game. Parrot wants to turn it into a feed. Short-form video content built on Comprehensible Input methodology, with power users spending 6 hours a week on the platform. The founders tested 52 language apps before building their own.
A million data points per game. Automatic play tagging. A transfer portal database with 2,400 players. PlayVision is turning game film into searchable, queryable intelligence for basketball programs that cannot afford a full analytics department.
Factory quality inspection has been stuck between expensive custom vision systems and generic models that do not understand production lines. Allus AI built a billion-parameter foundation model trained on 1.5 billion industrial data pairs, and it is already running on five Fortune 500 production lines with 99.95 percent accuracy.
Commercial insurance brokerage is a $100 billion industry where producers spend more time filling out forms than closing deals. Casey automates the entire submission process, from intake to carrier applications, so brokers can focus on what actually makes money: relationships.
Product demos are the bottleneck in every SaaS sales funnel. Primer built an AI agent that delivers personalized walkthroughs 24/7, and Figma, Gamma, and Lovable are already using it. That customer list tells you this is not vaporware.
Everyone is racing to generate code faster. Scott AI is the rare product that wants you to slow down first. Spec before codegen. Debate before commit. It sounds counterintuitive until you realize how many production fires start with 'nobody reviewed the architecture.'
Seventy percent of Fortune 500 companies still run mission-critical systems on COBOL. The engineers who built them are retiring. Hypercubic is racing to capture their knowledge before it walks out the door.
Nursing facilities waste billions on lost, broken, and over-rented medical equipment every year. Norra uses physical trackers and AI agents to fix the problem, and one facility already cut rental costs by 70 percent.
Enterprise codebases are too big for anyone to understand. Sourcebot gives both humans and AI agents the ability to search, query, and reason across millions of lines of code without sending anything off-premises.
Every parent wants their kid to spend less time gaming. Aleph Lab is betting on the opposite: what if the gaming itself was the classroom? Their AI buddy teaches kids English inside Minecraft, and the linguistics research says it might actually work.
The accounting industry is a $900 billion market built on rules and judgment calls. Cranston AI thinks both of those things can be automated. Their AI coworker plugs into QuickBooks, NetSuite, and Xero to handle reconciliation, tax compliance, and financial analysis. The question is whether accountants are a bottleneck or a moat.
Healthcare billing teams spend most of their day on hold with insurance companies, checking claim statuses one by one. LunaBill built AI voice agents that make those calls for them. $764K in contracted ARR, 50,000+ calls automated, and a 100% pilot-to-paying conversion rate. This is not a demo. This is a business.
Building an AI agent is getting easier. Deploying one into production, with persistence, scaling, safety guardrails, and monitoring, is still a multi-week engineering project. Arcten wants to compress that into minutes.
Sales reps lose deals when they cannot answer technical questions on the spot. Aside listens to live calls and surfaces answers from docs, Slack, and past winning calls in under a second.
Every AI company needs human annotators but finding, vetting, and managing them is a nightmare. Fixpoint automates the entire pipeline from sourcing to quality control.
Job applications are broken. Sorce lets you swipe on jobs like dating profiles and sends an AI agent to fill out the applications for you. Over 700K users and 20 million swipes say the idea works.
Industrial distribution runs on faxes, phone calls, and sales reps manually keying quotes into ERPs that were built in the 1990s. Avent automates quoting and order entry so human sales teams can stop doing data entry and start doing relationship management.
The DAW market has not had a real challenger in over a decade. GoRiff thinks AI can crack it open by making production faster for pros and accessible for everyone else.
Pre-sales CRM is a solved problem. Post-sales is still a disaster of spreadsheets, tribal knowledge, and missed renewal signals. Paloma unifies contracts, consumption data, and customer context into one system that actually tells you which accounts are about to churn.
Firmware engineering is one of the last software disciplines untouched by AI. BootLoop changes that with an agent that reads datasheets, writes production-ready embedded code, and validates it against actual hardware using oscilloscopes and logic analyzers.
The sales intelligence market is dominated by Apollo, ZoomInfo, and LinkedIn Sales Navigator. Clodo is betting that AI-native lead discovery, where you describe your ideal customer in plain English, makes all of them feel like phone books.
Every developer has built a weekend project that works locally but would take three weeks to make production-ready. Modelence ships everything you need to skip the boilerplate and go live from day one.
Voice AI is eating the world, but the training data supply chain is still held together with duct tape and crowdsourced recordings of questionable provenance. Panels wants to fix that with 10,000 vetted contributors across 100+ countries.
Debt collection is a massive industry run on manual phone calls and bad software. Altur built voice AI agents that negotiate payment plans and follow up on WhatsApp. Six months in Mexico, $6 million collected.
Security awareness training has been stuck on fake phishing emails for a decade. GhostEye builds exposure maps of employees and launches autonomous simulated attacks across email, voice, SMS, and deepfakes.
The Pentagon has a $500 drone problem that costs $100K to solve. Perseus Defense is building mass-manufactured self-guided missiles that bring the cost of killing a drone below $10,000.
Granola and Otter record your meetings and send everything to the cloud. Char does the same thing but stores everything locally in plain markdown files you actually own. Seven thousand GitHub stars suggest people care about the difference.
AI agents fail in ways that are hard to reproduce and harder to diagnose. Fulcrum builds red-teaming agents that find the failures for you, uncover environment bugs, and catch reward hacking before it ships.
Cursor and Copilot built AI coding for software engineers. Lotas built it for data scientists who live in RStudio, write R, and have been completely ignored by the coding assistant wave.
Biology labs spend more time wrestling with data visualization than doing actual science. Novaflow takes raw experimental data, a plain English question, and returns publication-ready plots in minutes.
Traditional pharma takes a decade and $2 billion to bring a drug to market. Convexia uses AI agents to source overlooked drug candidates, run computational trials, and move 10x faster than incumbents.
Multifamily property managers spend tens of thousands a year on manual expense coding. IronLedger automates the entire accounts payable workflow with AI agents that hit 97% accuracy out of the box and integrate with every major PMS.
B2B sales teams spend hours stitching together data from ZoomInfo, LinkedIn, and a dozen other sources. Orange Slice replaces the entire enrichment workflow with a spreadsheet where every column runs AI-generated TypeScript, pulling from 100-plus data sources in one pass.
Motion graphics have been trapped behind After Effects for two decades. Hera replaces the timeline with a text prompt and generated 50,000 animations in its first week.
Freight brokers spend their days making hundreds of phone calls to negotiate rates with carriers. Lanesurf built voice AI that handles those calls in parallel, and the system gets better at negotiation every time it dials.
Hospitals are drowning in AI vendor pitches and terrified of deploying any of them. Parachute gives health systems a governance layer to evaluate, monitor, and audit clinical AI without needing a team of ML engineers on staff.
Retail investors make up 25% of U.S. trading volume but more than half lack confidence in their decisions. Astor monitors markets continuously and delivers institutional-grade research briefs, catalyst alerts, and hypothesis testing to individual investors who want to stop guessing and start reasoning.
Running scientific simulations involves weeks of manual setup, parameter tuning, and validation. Fluidize lets researchers describe what they want in plain language and turns it into reproducible computational pipelines that scale from laptop to cloud.
AI coding tools are powerful but disconnected from what you actually see on screen. Stagewise puts a coding agent directly in your browser so you can click on elements and tell it what to change.
Property management companies lose leads and frustrate tenants because nobody answers the phone. Wayline puts an AI voice agent on every line, 24/7, and integrates with the property management software you already use.
Everyone wants to build AI agents that use computers like humans. The problem is you cannot train those agents on production systems without breaking things. Halluminate builds realistic sandbox environments that replicate Salesforce, Slack, and enterprise tools so AI labs can train and benchmark computer use agents safely.
Car washes are a $15 billion industry running on software from the early 2000s. Nautilus plugs into legacy POS systems and layers on AI-powered CRM, e-commerce, and marketing automation to help operators sell more memberships, cut churn, and stop managing their businesses out of spreadsheets.
Most companies are drowning in repetitive tasks that are too complex for simple automation but too tedious for humans. Trace sits in the middle, routing work to AI agents or people based on what actually requires human judgment.
Dental offices spend thousands each month ordering supplies from multiple vendors with no easy way to compare prices. Alara aggregates suppliers into a single cart with real-time price comparisons.
Legacy Windows applications run the backend of healthcare, finance, and logistics. Nobody builds integrations for them because nobody wants to. Cyberdesk trained an AI agent that learns how to use these apps the same way a human would, and hospitals are already using it for patient intake.
AI coding agents are powerful but chained to your laptop. Omnara cuts the cord by letting you monitor, review, and control Claude Code and Codex from your phone or any browser.
Commercial real estate brokers work 60-hour weeks and spend more than half that time on repetitive deal materials. Closera, built by two Stanford CS grads with backgrounds at Google and BCG, automates the grunt work of CRE so brokers can focus on relationships and deals.
AI models can pass bar exams but still produce UIs with white text on white backgrounds. Design Arena pits AI models against each other on real creative tasks and lets actual humans vote on which output is better. Simple concept, hard to replicate, and increasingly important.
Most coding agents lose the plot halfway through a long task. Magnitude splits the work across specialized subagents for planning, reviewing, debugging, and browsing so the main agent stays focused on what you actually asked for.
Two founders forked Chromium's 30-million-line codebase and built a browser where AI agents handle your tedious web tasks. Comparing flight prices, scheduling meetings, filling out forms. Meteor wants to be what Chrome would look like if it were designed for the agent era.
Recruiting is broken in a specific way: the first interview is almost always a waste of time for everyone involved. Outrove built an AI recruiter with a human-appearing video presence that joins Google Meet and Zoom calls, conducts natural screening interviews, and feeds structured evaluations into your hiring pipeline. Two founders who previously built seven-figure SaaS companies and a UN-recognized robotic doctor think the first 30 minutes of every hiring process should be automated.
Deel charges $599 per employee per month for EOR services. Remote charges $599. Rippling is somewhere in that range. Shor is launching at $99 per month for full-time employees and $20 per month for contractors, using AI agents for compliance automation and stablecoins for instant settlement. A Waterloo and UT Austin team thinks the global payroll industry is overcharging by 5x.
Fraud teams at fintechs and banks are drowning in false positive alerts while real fraud slips through. Socratix AI, built by a DoorDash fraud systems veteran and a Cruise infrastructure engineer, deploys autonomous AI analysts that investigate alerts around the clock and actually explain their decisions.
Self-insured employers waste over $60 billion a year on medical claims that should never have been paid. Avelis Health uses machine learning to review every single claim, voice AI to retrieve medical records, and LLMs to validate whether the billed services actually happened. They have already helped 60 patients erase over $300K in bills.
Most BI tools demand weeks of setup, SQL fluency, and a dedicated analyst to get anything useful out of them. Pleom connects your data sources in seconds and generates AI-powered visualizations, workflows, and analytics with zero learning curve. Founded by an 18-year-old Georgia Tech ML grad, it is the anti-Tableau.
Production debugging is still mostly manual. An engineer gets paged, opens Datadog, stares at logs, traces the issue across three services, and writes a fix four hours later. TraceRoot built an open-source AI agent that connects to your telemetry, traces the root cause, and drafts the pull request. The SDK has over 10,000 downloads and the founding team fixed 300+ production bugs at Meta and AWS before deciding to automate themselves out of a job.
GitHub has become the default. That is the problem. gitdot is not building a GitHub clone with more features. They are building an opinionated platform for open-source maintainers who care about code quality over vanity metrics. No AI copilot. No free private repos. No stars. A Rust-powered Git server and a CI/CD platform designed for people who treat code as craft.
Half of all search volume could shift to AI interfaces by 2028. That is a $1.3 trillion advertising opportunity with no incumbent ad network serving it. Imprezia, built by a former Amazon Ads and Meta Ads engineer, is placing contextual brand integrations inside AI chatbot conversations. One-line SDK. No banner ads. Just relevant brand mentions woven into natural dialogue.
Companies are deploying AI tools faster than their compliance teams can review them. Truth Systems monitors every AI interaction in real time, blocks non-compliant prompts before they process, and builds immutable audit trails. It is essentially a firewall for AI behavior, and they are starting with law firms.
Industrial robots take months to program for new tasks. Verne Robotics uses diffusion models to train bimanual robot arms from thirty minutes of human demonstration. Their first product deploys in four days. A biotech unicorn is already a customer. Two founders from Stanford, Columbia, and the Vision Pro team are building the learning layer that makes robots actually flexible.
Spreadsheets have not changed fundamentally since VisiCalc. BitBoard is building an AI-native version where you talk to an agent that builds, updates, and analyzes your data while you keep full control over every cell. Think of it as the spreadsheet rebuilt for people who would rather describe what they want than write formulas.
Three foreign producers control 97% of global magnesium and 96% of titanium sponge. Duranium is building a reactor in Alameda that recycles its own emissions and claims it can hit Chinese pricing without subsidies. That is a bold claim. The chemistry behind it is surprisingly credible.
NYC rental commissions are absurdly high, and landlords have been paying 12%+ because the alternative was doing everything themselves. Spotlight Realty built an AI-powered brokerage that handles tenant screening, showing scheduling, and the entire listing process for 4%. The agent never sleeps, and it never asks for a bigger cut.
End-to-end testing is the part of software development that everyone agrees is important and nobody wants to do. Docket replaces brittle CSS selectors with coordinate-based AI agents that see your app the way a human does, and the tests actually stay green.
Financial services firms spend billions on manual operations that follow rigid, documented procedures. Eloquent AI watches how your team works, learns the playbook, and then runs it autonomously. They hit $500K in annual recurring revenue in their first month, which tells you exactly how desperate this industry is for automation.
Everyone uses ChatGPT to write now. Almost nobody is happy with how it sounds. TypeOS is betting that the real product isn't better generation but better translation from AI-speak to your voice, and 6,000 users seem to agree.
Auth0, Clerk, and Supabase Auth all want to host your authentication. Better Auth thinks you should own it yourself. With 13K GitHub stars and 350K monthly npm downloads, developers seem to agree.
Computer-use agents need somewhere to work that is not your laptop. Cua provides sandboxed cloud desktops where AI agents can see screens, click buttons, and run applications without touching your actual machine.
A former Anthropic and Scale AI engineer is building a consumer social product powered by AI. The website says almost nothing. The investor list says plenty.
Cloud gaming promised instant play and delivered input lag. Wavedash runs games natively in your browser using WebAssembly and WebGPU, and it charges developers 10% instead of Steam's 30%.
An ex-AWS Bedrock engineer and an ex-Microsoft Copilot researcher built the testing infrastructure that every company deploying conversational AI needs but nobody has built yet.
Data centers are expanding faster than operators can staff them. Boost Robotics thinks autonomous robots that can actually manipulate hardware are the only way to keep up with the inspection backlog.
Industrial robots cost six figures and need expert programmers. The Robot Learning Company thinks dual-arm robots under $10,000 with self-learning software can change that math entirely.
Describe the prediction you want to make, connect your data, and Plexe builds, evaluates, and deploys the model for you. That is either the future of ML or a very good demo.
Everyone agrees referral marketing works. Almost nobody builds a referral program. Refgrow thinks the gap is friction, and it might be right.
Every AI company promises your data is safe. Tinfoil skips the promise and gives you a mathematical proof instead.
India's fashion and lifestyle market is racing past $100 billion, but most global brands still stumble at the border. Confluxe just raised $1.6 million to build the tech-led operating layer that gets them in without losing what makes them distinctive.
The founders of viafintech sold their last prepaid company for nine figures. Now they're back in Berlin, building the API layer they wish had existed the first time around. Finperks aggregates 700+ gift card and prepaid brands into a single integration, and they just raised $4 million to prove the plumbing matters more than the storefront.
CEOs spend half their week chasing status updates across Slack, Jira, and Notion. Bond claims it can do that job automatically and give them 10 hours back. The pitch is clean. The execution is what matters.
The global second-hand golf equipment market is worth $3 billion, but buying used clubs still feels like gambling on eBay circa 2008. Edinburgh-based Hosel just raised £500K to aggregate trusted sellers and bring recommerce infrastructure to a sport that desperately needs it.
Three PhDs from EPFL and Meta AI are building an AI that learns your company's institutional knowledge, runs experiments, and ships programs on your existing stack. The ambition is large. The question is execution.
The ocean is the last major physical infrastructure not governed by software. Mirai Robotics, founded by the team behind Blackshape Aircraft, Sketchin, and Talent Garden, just raised $4.2 million to build autonomous surface vessels for defense, surveillance, and maritime patrol. They're doing it from Puglia, not Palo Alto.
Every few months a new tool shows up promising to make data analysis accessible to people who don't speak SQL, and every few months I have to figure out if this one is actually different.
Every founder I know has a backlog of questions their engineers haven't had time to answer yet. Dex is betting that's a product, not a scheduling problem.
Two brothers, one from SpaceX and one from a clean energy startup, are using AI to match complex weld jobs to the right fabrication shops. The pitch: cut bid and scheduling time by 80%.
Construction runs on WhatsApp. Scopey Onsite uses AI to convert the thousands of messages, photos, and voice notes flying around every job site into structured, time-stamped records that hold up in disputes. They just raised $850K to prove the industry will pay for evidence it already creates.
Every team I know is drowning in Slack threads that mean nothing and missing the ones that matter.
The notch on your MacBook is useless dead space, and someone finally did something interesting with it.
Clidey started with WhoDB, an open-source platform for modeling organizations. Now they have added Docucod, which generates and maintains documentation straight from your codebase. Two products, one team, and a bet that data-driven decision-making needs better infrastructure.
The 'Cursor for documents' pitch is either the most accurate analogy I've heard this year or the most convenient one.
Hard tech companies need custom parts fast. Foundation Industries thinks the supply chain for low-volume, high-mix manufacturing is broken enough to rebuild from scratch.
AI product photography is a crowded, slightly overhyped corner of e-commerce tools right now, and SellShots knows it.
Every B2B SaaS app eventually needs comments, presence indicators, and notifications. Velt thinks building those from scratch is a waste of your engineering team's time.
Data centers run on five different software systems that don't talk to each other. Aravolta built one platform that replaces all of them and deploys in 48 hours.
The supply chain attack problem is real, unglamorous, and quietly getting worse — and Koidex just built a search bar for it.
Startups are paying enterprise prices for live chat tools they use like glorified sticky notes, and Foxchat thinks the fix is simpler than anyone's admitting.
Engineering teams spend 20 hours waiting for code reviews. Greptile indexes the whole codebase and cuts that to under 2. Over 1,000 teams already use it.
Every AI writing tool I've tested still makes me leave whatever I'm doing. TypeBoost is betting that's the whole problem.
VibePad maps your PlayStation controller to Claude Code and Codex shortcuts, and I'm annoyed at how much sense it makes.
The team that built Gatsby.js is back with an open-source TypeScript framework for AI agents. 22,000 GitHub stars suggest developers are paying attention.
Every AI assistant claims to save you time. Tidy is the first one I've seen that admits it needs you to teach it first, and somehow that's the most honest pitch I've heard all year.
Every product team has a graveyard of feedback tickets that say almost nothing useful, and Woise thinks the fix is letting people just talk.
Apple Watch hardware is genuinely good — it's the software leash that forces you to buy the whole ecosystem. Merge is betting you're tired of that deal.
Your data pipelines are broken more often than you think. Acceldata built an observability platform that monitors data quality, pipeline health, and infrastructure costs so you find out before your CFO does.
A new AI marketing platform claims it can read your code and run your growth strategy — which is either exactly what solo founders need or a very confident promise.
An SEO agency got tired of GA4 lying to their face, so they built the thing that actually tells you who's crawling your site — and it turns out, it's mostly not humans.
There are hundreds of meditation apps. Most of them are glorified audio players with a subscription model. Sati AI is trying something different: a personalized AI teacher grounded in Buddhist psychology and neuroscience. Whether that distinction matters is the interesting question.
A Chrome extension that summarizes the legal fine print you've been clicking through for twenty years — sensible idea, crowded-ish space, 105 users so far.
Postgres scaling has always meant painful tradeoffs. PgDog, built by the engineer who sharded Postgres at Instacart during peak pandemic growth, wants to make horizontal scaling something you drop in rather than something you dread.
Another AI sales tool launched on Product Hunt this week — except this one hit #1 daily with 558 votes, which means it's worth actually looking at.
Subscription Day² just landed on iOS with 307 upvotes and a calendar-first take on a problem everyone has but nobody loves solving.
Unicorne shows you the 20 fastest-growing startups in real time, with verified revenue data — which is either the most useful dashboard you didn't know you needed, or a beautifully designed distraction.
Embedding questions directly into the video timeline sounds obvious in retrospect — which is either a good sign or a warning.
Writing a defense proposal takes months, costs hundreds of thousands of dollars, and still might lose. ConductorAI built an AI platform that searches millions of government documents to help contractors write compliant proposals faster.
Everyone knows spray-and-pray outreach is dead. Fewer people have a credible alternative. Extrovert thinks the answer is commenting.
SaaS pricing got complicated. Hyperline built a quote-to-cash platform that handles usage-based billing, invoicing across 100+ countries, and the entire revenue lifecycle without requiring a six-month Salesforce implementation.
The Mac mini M4 became the accidental hardware of the agentic AI moment, and now someone made it a cute crustacean costume.
SQLite for the web is a solved problem with about five solutions. Bunny Database is betting its CDN roots make it the one that actually sticks.
Campfire has quietly built the most comprehensive creative writing platform on the market. 750,000 authors, 4 million projects, and a worldbuilding toolkit that makes Scrivener look like Notepad.
Turning Google Sheets into a REST API isn't a new idea, but Sheetful is betting there's still a real market for doing it cleanly and for free.
Everyone wants an AI agent that texts them back on WhatsApp. Almost nobody wants to spend a weekend debugging Python environments to get there.
The Play Store's video compressor category is a swamp of ads and paywalls — and one open-source Kotlin project just walked in and made everyone look slow.
Every AI agent needs to talk to the outside world. Klavis AI is building the managed infrastructure layer so developers don't have to wire up OAuth flows and tool integrations from scratch.
The job market is soft, ATS systems are broken, and a Carnegie Mellon-connected team thinks structured project portfolios are the answer — at least for tech candidates who can actually show their work.
Most AI support bots hallucinate answers your docs never contained. Inkeep grounds everything in your actual content and then lets you build agents on top of it. The customer list suggests it works.
Getting to the Moon is expensive because every mission needs its own dedicated rocket. Instinct is building small lunar landers designed to launch as regular LEO rideshare payloads, cutting costs and increasing how often anyone can actually get there.
Kafka is powerful, popular, and notoriously difficult to work with. Quix built a platform that lets engineering and data teams develop, deploy, and monitor stream processing pipelines without wanting to throw their laptops out the window.
Construction estimating is still done by humans squinting at blueprints and punching numbers into spreadsheets. Bild AI wants to automate the entire process from blueprint to cost breakdown.
Every developer building AI agents is solving the same infrastructure problems from scratch. Truffle wants to hand them a solved stack so they can focus on the actual agent logic.
Traditional recruiting agencies charge 25% of first-year salary and still miss half the good candidates. Candid embeds AI talent engineers into your hiring process for $500 a month per role. The math speaks for itself.
Two teenage founders built an AI grading tool to $200K ARR before they could vote. Now they're scaling it across Australian schools with Y Combinator backing.
Hotels field thousands of calls a week and most of those calls are the same twenty questions. Flowtel is building AI agents that handle everything from booking to room service.
The average sales team uses six different tools to do prospecting. Verbiflow thinks one AI agent can replace the entire stack. They might be right.
Two Stanford engineers are compressing years of urban planning studies into months. Municipalities spend $50 billion a year on planning consultants. Waypoint wants a piece of that.
Upload a photo. Add a voice. Get a real-time AI avatar you can have actual conversations with. Dollyglot is building Character.AI with faces, and the tech underneath is genuinely impressive.
AI image generation has a usability wall. Lightbox is betting that pre-built pipelines and a conversational interface can knock it down for the rest of us.
The old coding interview tests whether you can solve algorithms on a whiteboard. Outship tests something different: can you actually ship software with AI tools? That might be a better signal for how engineering actually works now.
Live voice translation has been a sci-fi promise for decades. Pinch is building the API and desktop tools to make it work in actual meetings, with actual tone preservation.
Wiz sold to Google for $32 billion. SubImage thinks the same problem can be solved with open-source foundations and without the vendor lock-in. They might be right.
Customs classification is one of those problems that sounds boring until you realize companies are bleeding billions a year getting it wrong. Trava wants to fix that with AI agents that audit entries and automate tariff codes.
Enterprise search has been mediocre for decades. ZeroEntropy is building an AI-native search API that handles the messy, unstructured documents most RAG systems choke on.
Every sales team hates updating their CRM. Ergo monitors Zoom, email, and Slack to keep HubSpot and Salesforce current automatically, then goes further by executing follow-ups based on market signals.
Documentation drift is the silent killer of developer experience. Promptless monitors your PRs, Slack threads, and support tickets, then updates your docs automatically. No prompting required.
Every company says they need better documentation. Nobody wants to write it. Peppr AI watches your Slack, Notion, and Jira, then writes the docs for you. The pitch is compelling. The graveyard of knowledge management startups is also very large.
Most ed-tech products are built by people who have never taught a class. Alice's CEO spent eleven years in university classrooms, and it shows. 100K students, $20K MRR, and a product that turns your messy PDFs into actual exam prep.
Most AI video models generate clips you watch. Lucid built one you interact with, running at 20+ fps on a single GPU with 100x less training compute than the competition.
Month-end close is the accounting world's recurring nightmare. Mesh automates 90% of accruals by pulling real-time signals from AP inboxes, Slack, and Teams, promising to shave days off the close cycle.
More than a billion people in emerging markets watch their savings erode against the dollar every year. Karsa wants to hand them a stablecoin-backed USD account, a Visa card, and 4-5% yield, all from a single app.
Vintage shopping has a discovery problem. Retrofit built an AI personal shopper named Sophie to fix it, pulling from independent stores and resellers across the country.
Every company has a graveyard of SQL scripts, cron jobs, and one-off connectors that nobody wants to touch but everything depends on. Superglue replaces them with AI-powered, self-healing integrations.
Commercial insurance brokerages run on email, PDFs, and spreadsheets. Vantel is building an AI-native OS that handles the document processing, CRM, and workflows that brokers currently do by hand.
Grid-scale battery storage is a fast-growing, data-heavy market where operators are making million-dollar dispatch decisions with spreadsheets and stale reports. Dartboard Energy wants to replace the analyst, not the dashboard.
Several top-20 hedge funds are already using Finbar to automate financial modeling, research, and data analysis. When your customers manage hundreds of billions in assets, the margin for error is zero and the tolerance for hype is lower.
Outside the prime lending market, incumbents charge predatory rates because they can. Karoo is building an AI-native lending platform that underwrites better and prices fairer for the millions of Americans stuck between big bank rejection and payday loan traps.
A former Citadel energy trader figured out that grid congestion patterns are predictable, and now he's using AI to buy land for solar and wind projects before the competition even knows where to look.
Half of all Medicare members churn annually because agencies can't keep up with post-enrollment engagement. careCycle deploys multi-agent AI voice teams that handle the entire customer journey, claiming 5x higher conversions and 37% better retention. The numbers are big. The problem is real.
Running transformer models on edge devices usually means accepting terrible performance or buying expensive hardware. Exla's SDK aggressively quantizes models for up to 80% memory reduction and 3-20x faster inference with a few lines of code.
Software engineers got GitHub, VS Code, and Cursor. Hardware engineers got... spreadsheets and PDFs. Artifact is building a collaborative IDE for electrical systems design, and they've already got Boom Supersonic and K2 Space as customers. This team built avionics for hypersonic aircraft. They know exactly what's broken.
ASML has a near-monopoly on extreme ultraviolet lithography machines. Inversion Semiconductor wants to build something fundamentally different by shrinking particle accelerators 1000x. Two guys from Imperial College and UCL with backgrounds in rockets and CERN. The ambition level here is genuinely staggering.
Design tools and code editors have been separate universes for decades. Onlook merges them into one visual editor that writes directly to your React codebase. With 22,000 GitHub stars and growing, this might be the tool that finally kills the design-to-dev handoff.
Trata uses AI agents to interview hedge fund analysts, synthesize investment research, and distribute it as a subscription product. They already have 2,000+ tickers and 125+ fund contributors. The question is whether fund managers will trust research they didn't pay a human to write.
Roofing contractors lose thousands on every job because their estimates are sloppy. Maive uses computer vision and AI to fix the estimates before the crew shows up, and the founders have 15 patents and a unicorn exit between them.
GPU compute is the new commodity. Pluto is building a regulated derivatives exchange where you can hedge against compute price swings the same way airlines hedge jet fuel. The question is whether the market is ready to trade compute like oil.
What happens when you buy profitable software businesses and replace the human teams with AI agents? Rocketable is running that experiment with real money, and the thesis is more interesting than the headline suggests.
Law firms spend thousands on ads to get the phone to ring, then lose half those leads to voicemail and slow follow-up. Caseflood.ai wants to be the AI layer that catches every call and converts it before the prospect moves on.
Pharma companies spend millions finding the right doctors to consult on drug development. G LNK thinks AI matching can replace the Rolodex-and-phone-call model that expert networks have relied on for decades.
